A question to consider, that we have been asked (and asked ourselves) recently:
‘As a data science and analytics company, what role should social media play in the way you engage with customers, clients, or anyone in your circles?’
Does a company like ours belong on a channel like Instagram?
Take this example, where we use an R script to extract the RGB values of a JPEG, convert these values into a CSV, and then pass that CSV file through a K-means clustering algorithm on Microsoft Azure and regenerating the image. The Azure Machine Learning portal already contains their own example of this with a Bill Gates RGB, but you’ll never know when you need to map the RGB co-ordinates of your own image(?)
library(reshape2) # install.packages("reshape2")
path <- "/path/to/file/"
test <- readJPEG("image.jpg") # read in JPG as co-ordinates
r = melt(test[,,1]*255) # convert to RGB values
g = melt(test[,,2]*255)
b = melt(test[,,3]*255)
# create a file format readable by Azure ML
rgb.data <- data.frame(X = r$Var1, Y = r$Var2, R = r$value, G = g$value, B = b$value)
write.csv(rgb.data, "rgb.csv", sep = ",", row.names = F, quote = F) # create file
Here’s the before image:
And the after image:
We have essentially mimicked the common Instagram filter, but by using a clustering algorithm to group values of RGB co-ordinates together based on similarity, that when they are plotted will produced a cool colour image.
Does a company like ours belong on Twitter?
We have previously used YouTube to showcase our Data Science Portal (which is in need of a bit of love and attention) but in the run up to SQL Bits, we want to contribute new content to all communities relating to data and data science. For instance, we recently explored the topic of automation in industry and whether we are indeed going to lose our jobs to machines. Lachlan McLachlan explains:
This one is for fun.
We are attending SQL Bits XV in Liverpool this year from the 4th-7th May 2016. We would love you to follow our antics – search for Consolidata and follow our story as it happens.
Asking the question of ‘do we need this channel?’ ensures that we do not fall into the trap of ‘jumping on the bandwagon’ with regards to social media engagement. Saving time by:
- Choosing the best channels for engaging with other clients, prospects, data scientists and followers.
- Not replicating content across channels.
- Having the data available to us from our social media activities to guide us in the right direction…
… helps us to improve the way we engage with others. What do you think?